httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Aram Mirzadeh <>
Subject Re: Timeout patch uploaded
Date Mon, 22 Jan 1996 03:25:22 GMT
> Hi,
> 	just uploaded:
> for_Apache_1.1b0/84.timeout.patch

At least someone else is listening... if this isn't fixed soon I'm 
going to have to switch back to ncsa.  The amount of hits that my
main server is getting is increasing, for several reasons, one being
my my apache pages.  This is creating more and more timeouts in my
error log file, which means more and more people are getting incomplete

I tried the following patch, it does reduce the number of accept failed
error, but they're still showing up.  +2 for the attempt, -1 for the 
fix though.

> 	From: (Nathan Schrenk)
> 	Subject: Timeout code fix
> 	Affects: http_protocol.c, http_main.c
> 	ChangeLog: Reset timeout timer after each successful fwrite()
> 		to the network.  This patch makes timeout_name a
> 		global, and modifies the send_fd() procedure to
> 		check for an existing timeout by looking at
> 		timeout_name.  If a timeout is set, send_fd() resets
> 		the alarm timer to the default timeout every time
> 		fwrite() successfully sends data.
> This is by far the most common bug reported against apache at the
> moment.  It's pretty useless as a webserver unless the problem is
> fixed.
> Cheers,
> Ay.

#define BITCOUNT(x)     (((BX_(x)+(BX_(x)>>4)) & 0x0F0F0F0F) % 255)
#define  BX_(x)         ((x) - (((x)>>1)&0x77777777)                    \
                             - (((x)>>2)&0x33333333)                    \
                             - (((x)>>3)&0x11111111))

                -- really weird C code to count the number of bits in a word
Aram Mirzadeh

View raw message